PROJECT SUMMARY/ABSTRACT Antibiotics are life-saving, but their suboptimal use can lead to antibiotic-resistant bacteria, which contribute to approximately 2,000,000 illnesses and 23,000 deaths in the United States annually. Unfortunately, the development of new antibiotics is currently quite limited. Thus, immediate solutions are needed to preserve existing therapies, in particular anti-pseudomonal beta-lactams, which are the backbone of treatment protocols in the intensive care unit (ICU). A >500-fold intra-individual variability in drug levels has been observed using our current standard of ?one size fits all? dosing extrapolated from non-critically ill patients. In 30%-50% of ICU patients beta-lactam levels are too low; this increases treatment failure 1.5-fold and propagates antibiotic resistance. Therefore, the overall objective of this proposal is to develop a highly predictive, user-friendly, individualized beta-lactam dosing and monitoring tool for use in the ICU. This will be accomplished through three specific aims: (1) Design accurate individualized up-front beta-lactam dosing models for critically ill patients, (2) Develop and validate a dynamic prediction model for use during therapy that identifies patients at highest risk of either treatment failure/new resistance (drug level potentially too low) or neurotoxicity (drug level potentially too high), and (3) Identify factors that enhance implementation of these individualized dosing and monitoring models in practice. For Aim 1, we will prospectively assay beta-lactam levels from residual blood samples from 300 ICU patients and use population pharmacokinetics to develop a novel individualized dosing tool. For Aim 2, we will query the electronic records of 5,000 adult ICU patients treated with anti-pseudomonal beta-lactams to develop and validate a highly-predictive model that classifies patients as either low-risk (high likelihood of a favorable outcome, no drug level monitoring indicated) or high-risk (high likelihood of an unfavorable outcome, drug level testing may be beneficial). Beta-lactam use is nearly ubiquitous in critically ill patients. To optimize the potential for future clinical translation, in Aim 3 we will use focused ethnography informed by implementation science to identify factors that influence implementation of these new tools in trials and real-world practice. Dosing and monitoring protocols and interfaces will be iteratively refined based on these insights. The proposed Career Development Award supports the NIAID mission by striving to improve the care of patients with infections while providing the research skills, training, and mentorship necessary to develop the candidate into an independent investigator conducting patient-oriented research. The proposed training in pharmacokinetic modeling, bioinformatics, and dissemination and implementation science coupled with an outstanding multidisciplinary mentorship team, and the exceptional resources available at Mayo Clinic will allow the candidate to achieve her long-term goal of becoming an independent clinical researcher focused on improving the health of critically ill patients with infections.